AI Implementation
Article
Developing the AI Healthcare Resource Center began as a course project, but over time it evolved into something larger: a professional resource intended to help healthcare educators, simulationists, and organizational leaders move from curiosity about artificial intelligence to responsible implementation.
That evolution has required me to think beyond whether the prototype is useful today. I must also ask whether it can grow, whether it can be sustained, and whether its continued use will remain ethical as artificial intelligence, organizational policies, and user needs change.
In health professions education, a successful pilot is not the same as a sustainable innovation. A resource may initially attract attention, receive positive feedback, and demonstrate value, yet still disappear when the original champion moves on, the technology changes, or competing priorities emerge. For this reason, the future of the AI Healthcare Resource Center depends not only on expanding its content, but also on embedding it into structures, relationships, and practices that allow it to outlast the course and, eventually, its original creator.
The Innovation
The AI Healthcare Resource Center is a digital platform designed to support the responsible implementation of artificial intelligence in healthcare education and simulation. It includes educational articles, implementation frameworks, accessibility resources, examples of AI-enhanced teaching strategies, podcasts, curated references, and practical guidance for educators who may be interested in AI but uncertain about where to begin.
The core problem the website addresses is not a lack of information about AI. Information is abundant. The more significant problem is that healthcare educators often lack practical, trustworthy, and context-sensitive guidance for moving from awareness to implementation.
The website is therefore not intended to function as a list of tools. Its purpose is to help educators ask better questions:
Is AI appropriate for this educational problem?
What organizational conditions must be in place before implementation?
What ethical, privacy, accessibility, and equity concerns must be addressed?
How should AI-generated content be reviewed?
How can an innovation be evaluated, adapted, and sustained over time?
Can the Innovation Scale?
The AI Healthcare Resource Center has the potential to scale because it is digital, modular, and accessible beyond a single classroom or institution. Unlike a face-to-face workshop that requires repeated delivery, a web-based resource can reach additional users without a proportional increase in personnel or cost.
However, technical reach alone does not equal meaningful scalability. A website can be available globally and still fail to influence practice. For the innovation to scale effectively, users must find the content relevant, understandable, trustworthy, and applicable to their own contexts.
Several design features support scalability:
Modular sections that can be expanded independently
Content written for users with varying levels of AI literacy
Multiple formats, including articles, podcasts, visual resources, and downloadable tools
Resources that can be used by individuals, faculty groups, and organizations
A focus on implementation processes rather than a single technology platform
The ability to incorporate new scholarship, case studies, and emerging guidance over time
At the same time, scalability introduces risks. As the audience expands, the website may be used by individuals working under different institutional policies, legal requirements, technological infrastructures, and cultural expectations. Content that is appropriate in one organization may not be appropriate in another. Scaling therefore requires clear language that encourages users to adapt resources locally rather than treating the website as universal policy or prescriptive guidance.
Diffusion of Innovation as a Lens for Growth
Rogers’ Diffusion of Innovation theory provides a useful framework for considering how the resource might spread. Rogers identified five characteristics that influence adoption: relative advantage, compatibility, complexity, trialability, and observability.
Relative Advantage
The resource must offer a clear advantage over general web searches or uncurated lists of AI tools. Its relative advantage is its focus on healthcare education, simulation, implementation science, and responsible use. Users should be able to locate resources that are more relevant to their work than the broad, fragmented information available elsewhere.
Compatibility
The website is more likely to be adopted when its resources align with existing educational responsibilities, institutional values, accreditation expectations, and faculty development goals. Rather than asking educators to abandon their current instructional practices, the resource encourages them to consider how AI might support existing work.
Complexity
AI already feels complex to many educators. If the website adds another layer of technical language or complicated implementation processes, adoption will be limited. Reducing unnecessary complexity is therefore a core design priority. Content should provide clear entry points for beginners while still offering deeper resources for experienced users.
Trialability
Users should be able to experiment with small portions of the resource before committing to a larger implementation effort. A faculty member might begin with a prompt guide, an ethical checklist, a scenario-development example, or a short article. These low-risk entry points make it possible to test the value of the resource within a local context.
Observability
Adoption is more likely when users can see how others have applied the ideas successfully. Future versions of the website should include case studies, implementation stories, sample workflows, and lessons learned. These examples will make the outcomes of responsible AI implementation more visible and concrete.
Rogers’ adopter categories also remind me that the website cannot be designed only for innovators and early adopters. Those users may already be enthusiastic about AI and willing to tolerate uncertainty. The more difficult design challenge is serving the early and late majorities, who may need stronger evidence, simpler processes, institutional approval, and visible examples before engaging.
If these users do not adopt the resource, the answer should not simply be that they are resistant to change. It may indicate that the design has not adequately addressed their concerns, workload, context, or level of readiness.
The Hype Cycle and the Current AI Environment
The Gartner Hype Cycle provides another important lens because generative AI remains surrounded by rapidly shifting expectations. The Hype Cycle describes movement from an innovation trigger to a peak of inflated expectations, followed by disillusionment, gradual learning, and eventual productivity.
AI in health professions education appears to be moving between heightened enthusiasm and growing concern. Institutions are simultaneously excited about efficiency and worried about privacy, bias, academic integrity, overreliance, and quality.
This environment creates both an opportunity and a risk for the AI Healthcare Resource Center.
The opportunity is that educators are actively seeking guidance. The risk is that the website could unintentionally reinforce hype by emphasizing possibilities without giving equal attention to limitations, evidence, and failed implementations.
For the resource to remain credible, it must not function as promotional material for AI. It must acknowledge uncertainty, distinguish evidence from speculation, and provide space for critical evaluation. It should also help users avoid two common mistakes: adopting AI too quickly because of enthusiasm or abandoning it entirely when early results are less transformative than expected.
The long-term value of the website will depend on helping educators move toward what the Hype Cycle describes as the plateau of productivity, where AI is used selectively, appropriately, and with realistic expectations.
Theory of Change
Theory of Change asks what must be true for the innovation to produce the outcomes I intend.
The intended long-term outcome is not simply increased website traffic. It is improved capacity among healthcare educators to evaluate and implement AI responsibly.
The proposed pathway is:
Accessible resources and practical implementation guidance
lead to
increased educator understanding and confidence
which supports
more thoughtful selection and testing of AI tools
which contributes to
safer, more equitable, and more sustainable implementation in healthcare education.
Several assumptions underlie this pathway:
Educators recognize a need for implementation guidance.
They trust the quality and credibility of the website.
They have time and institutional support to apply what they learn.
The resources are accessible and relevant to their context.
Institutional policies permit appropriate use of AI.
Users maintain human oversight rather than relying uncritically on AI outputs.
The website remains current as technologies and regulations change.
Making these assumptions visible is important because any one of them could fail. For example, increased educator knowledge will not automatically lead to responsible implementation if the organization lacks approved platforms, governance structures, leadership support, or faculty development time.
Theory of Change therefore helps clarify that the website is only one component of a larger implementation system. It can support readiness and capacity, but it cannot independently resolve organizational barriers.
Long-Term Sustainability
The website is potentially sustainable, but sustainability will require more than continuing to pay for hosting.
The greatest threat is dependence on a single champion. At present, much of the content development, organization, review, and maintenance depends on me. That is manageable during the prototype stage, but it creates vulnerability. If my professional responsibilities change, if competing priorities increase, or if the volume of content becomes too large, the resource could become outdated.
Several strategies could improve sustainability:
Embed the Resource Into Existing Work
The website aligns with my doctoral research, scholarly interests, professional role in healthcare simulation, and involvement in AI-focused collaborative work. Integrating website updates into these existing activities makes maintenance more feasible than treating it as a separate project.
For example, future conference presentations, publications, faculty development materials, implementation studies, and collaborative products can become new website content.
Establish a Content Review Cycle
AI content can become outdated quickly. Pages should include review dates, and high-risk topics such as privacy, regulation, tool recommendations, and institutional policy should be reviewed more frequently than stable theoretical content.
A structured schedule could include:
Quarterly review of emerging AI tools and guidance
Semiannual review of implementation resources
Annual review of site organization, accessibility, and user needs
Immediate review when major regulatory or policy changes occur
Develop Shared Ownership
Long-term sustainability would be strengthened through collaborators, contributors, or an advisory group. Shared ownership could reduce dependence on one person and broaden the range of expertise represented on the website.
Potential contributors might include simulation educators, instructional designers, clinicians, accessibility specialists, learners, data governance professionals, and implementation scientists.
Maintain Platform Flexibility
Dependence on a single commercial tool or web platform creates risk. Pricing, features, ownership, and access conditions may change. The content should therefore be organized in ways that allow migration, backup, and reuse across platforms.
Publish and Disseminate Findings
The resource will be more sustainable if its development and outcomes are documented through scholarly dissemination. Publishing findings, processes, and lessons learned allows others to build on the work even if the original platform changes.
Responsible Use at Scale
Scaling an innovation also scales its risks.
If a misleading recommendation appears on a small prototype, its impact may be limited. If the same recommendation reaches thousands of users, the potential harm grows significantly. Responsible scaling therefore requires stronger review processes, transparency, and accountability.
Bias and Equity
The website must acknowledge that AI tools may reflect biases in their training data, design, and evaluation. Resources should encourage users to examine whether AI outputs represent diverse populations and whether certain learners or patient groups could be disadvantaged.
The website itself must also avoid centering only well-resourced institutions with advanced technology. Responsible scaling means including examples and guidance that are relevant to smaller programs, limited-resource settings, and educators with different levels of technical expertise.
Accessibility
Scalability is not meaningful if the website is inaccessible. Content should be usable by individuals with disabilities, varied literacy levels, different devices, and inconsistent internet access.
Future development should include:
Routine accessibility testing
Alternative text for images
Captions and transcripts for multimedia
Clear headings and navigation
Readable formatting and contrast
Mobile-responsive design
Downloadable or low-bandwidth alternatives where possible
Cybersecurity and Data Governance
The website must clearly communicate that users should not enter protected health information, identifiable learner data, confidential institutional information, or proprietary materials into public AI systems without authorization.
As the resource grows, data collection through subscriptions, analytics, forms, or interactive tools must also be carefully governed. The website should collect only the information necessary for its function and clearly communicate how information is used.
Intellectual Property
The resource includes original writing, AI-assisted content development, scholarly references, and potentially contributed materials. Clear attribution, permission processes, and authorship expectations will be necessary as the website expands.
The website should also model responsible practice by distinguishing original work, adapted resources, licensed materials, and AI-assisted content.
Human Oversight
The resource must reinforce that AI should support rather than replace educator judgment. This is particularly important in healthcare simulation, where educational content can shape clinical reasoning, teamwork, and patient care.
Users should be encouraged to review AI-generated scenarios, assessment items, feedback, and recommendations for accuracy, bias, appropriateness, and alignment with learning objectives.
Costs and Trade-Offs
A digital platform may appear inexpensive, but sustaining it involves direct and indirect costs.
Direct costs may include hosting, domain registration, premium software, accessibility services, design support, cybersecurity, and content management tools.
Indirect costs include time spent reviewing evidence, updating content, responding to users, maintaining links, monitoring policy changes, and coordinating contributors.
There are also opportunity costs. Time spent developing the website is time not spent on other scholarship, teaching, operational work, or professional responsibilities. Sustainability therefore depends on ensuring that the resource remains closely aligned with my broader goals rather than becoming an isolated obligation.
AI use may reduce some development time, but it also creates additional review responsibilities. Faster content generation does not eliminate the need for verification. In some cases, it increases the need for careful oversight.
Environmental Sustainability
The environmental cost of AI must remain part of the discussion. Generative AI depends on energy-intensive infrastructure and substantial water use. Although a single website has a modest footprint, encouraging widespread and unnecessary AI use could contribute to a larger environmental burden.
Responsible design means avoiding AI use simply because it is available. The website should encourage purposeful use, efficient prompts, reuse of existing resources, and selection of the least resource-intensive tool that meets the educational need.
Environmental impact may be difficult to measure at the individual website level, but it should still be acknowledged as part of the innovation’s broader ethical footprint.
Monitoring Sustainability and Innovation Fatigue
An innovation may remain technically available while gradually losing relevance, trust, or use. Sustainability therefore requires monitoring more than whether the website is online.
Potential indicators include:
Reach
Number of unique visitors
Geographic and professional diversity of users
Repeat visits
Resource downloads
Newsletter subscriptions
Engagement
Time spent on key pages
Completion of podcasts or multimedia resources
Use of implementation tools
Participation in feedback opportunities
Requests for collaboration or presentation
Usefulness
User-reported relevance
Confidence before and after using resources
Examples of resources applied in practice
Testimonials or implementation case reports
Returning users who seek additional resources
Quality and Responsibility
Percentage of content reviewed on schedule
Broken or outdated links
Accessibility issues identified and corrected
Reported inaccuracies or concerns
Compliance with privacy and governance expectations
Innovation Fatigue
Declining engagement despite continued promotion
Reduced repeat visits
Low use of newly added features
Feedback indicating that the amount of content is overwhelming
Users reporting that AI guidance feels repetitive or disconnected from practice
Increasing maintenance effort without corresponding impact
Monitoring innovation fatigue is important because more content is not always better. If the website becomes too large, complex, or difficult to navigate, expansion could decrease its usefulness. Sustaining the innovation may therefore require removing outdated content, consolidating overlapping resources, and resisting the urge to add every new AI tool or trend.
What Must Remain Core and What Can Adapt
For an innovation to scale, it must maintain its core purpose while allowing adaptation.
The core elements of the AI Healthcare Resource Center are:
A focus on healthcare education and simulation
Responsible and ethical AI implementation
Human-centered design
Evidence-informed content
Implementation science
Accessibility and equity
Human oversight
These elements should remain stable.
Other features can adapt:
The specific technologies discussed
The platform used to host the resource
The content format
The organization of sections
The examples and case studies
The contributor model
The frequency of publication
This distinction between core and adaptable components will help the resource evolve without losing its identity.
Final Reflection
I believe the AI Healthcare Resource Center can scale and be sustained, but not automatically.
Its digital format, modular design, and alignment with a growing need in healthcare education create strong conditions for growth. Its connection to my doctoral research, professional practice, and scholarly interests also supports long-term maintenance.
However, the greatest risks are equally clear: dependence on a single champion, rapid content obsolescence, platform changes, insufficient governance, inaccessible design, and expansion without evidence of impact.
Diffusion of Innovation helps me understand how different users may encounter and adopt the resource. The Hype Cycle reminds me to avoid reinforcing unrealistic expectations about AI. Theory of Change makes the assumptions behind the innovation visible and clarifies that a website alone cannot create organizational change.
Most importantly, this reflection has reinforced that scalability should not be measured only by how many people can access the resource. Responsible scalability means expanding reach without diluting quality, increasing harm, or abandoning the values that shaped the original design.
Sustainability is not simply keeping the website active. It means building the structures, relationships, review processes, and evidence necessary for the resource to remain useful after the enthusiasm of the prototype stage has passed.
The goal is not to create the largest collection of AI resources. The goal is to create a trusted, evolving, and responsible resource that helps healthcare educators make better implementation decisions over time.
References
Center for Engaged Learning. (2024). AI, higher ed and the Hype Cycle. https://www.centerforengagedlearning.org/ai-higher-ed-and-the-hype-cycle/
Dubé, T. V. (2024). An imperative for transforming health professions education. Medical Education, 58(1), 8–10. https://doi.org/10.1111/medu.15274
Farrukh, K. (2024). Theory of change framework for program evaluation in health professional education. Pakistan Journal of Medical Sciences, 40(4), 793. https://doi.org/10.12669/pjms.40.4.8960
Gartner. (2024). Hype Cycle for healthcare providers, 2024. https://www.gartner.com/en/documents/5573627
Itani, A., Gronseth, S. L., Musaad, S., Nguyen, T., Mirabile, Y., & Beech, B. M. (2025). Ethical considerations for teaching with artificial intelligence: A scoping review in medical education settings. International Journal of Educational Technology in Higher Education, 22(1), 68. https://doi.org/10.1186/s41239-025-00563-9
Jhaveri, M., & Palat, V. (2025, June 27). Measuring and standardizing AI’s energy and environmental footprint to accurately assess impacts. Federation of American Scientists. https://fas.org/publication/measuring-and-standardizing-ais-energy-footprint/
Jose, B., Cherian, J., Verghis, A. M., Varghise, S. M., S., M., & Joseph, S. (2025). The cognitive paradox of AI in education: Between enhancement and erosion. Frontiers in Psychology, 16, 1550621. https://doi.org/10.3389/fpsyg.2025.1550621
Pham, T., et al. (2025). The impact of generative AI on health professional education: A systematic review in the context of student learning. Medical Education. https://doi.org/10.1111/medu.15746
Pusic, M. V., & Ellaway, R. H. (2024). Researching models of innovation and adoption in health professions education. Medical Education, 58(1), 164–170. https://doi.org/10.1111/medu.15161
Rogers, E. M. (1962). Diffusion of innovations. Free Press.